Ann Thorac Surg 2005;79:1704-1710
© 2005 The Society of Thoracic Surgeons
Original articles: General thoracic
Inflammatory Gene Polymorphisms Influence Risk of Postoperative Morbidity After Lung Resection
Andrew D. Shaw, MBa,*,
Ara A. Vaporciyan, MDb,
Xifeng Wu, MDc,
Terri M. King, PhDe,
Margaret R. Spitz, MDc,
Joe B. Putnam, MDf,
Burton F. Dickey, MDd
a Division of Anesthesiology and Critical Care Medicine, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
b Department of Thoracic Surgery, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
c Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
d Department of Pulmonary Medicine, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
e Department of Internal Medicine, University of Texas-Houston Medical School, Houston, Texas
f Department of Thoracic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee
Accepted for publication October 12, 2004.
* Address reprint requests to Dr Shaw, Dept of Critical Care Medicine, Unit 112, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030 (E-mail: ashaw{at}mdanderson.org).
 |
Abstract
|
|---|
BACKGROUND: Polymorphisms in genes encoding proteins involved in the inflammatory response may lead to a differential response to a noxious stimulus. We hypothesized that proinflammatory alleles at candidate loci would predispose patients undergoing lung resection to cardiopulmonary complications with a presumed inflammatory cause.
METHODS: We determined the genotypes at six candidate loci in 155 patients who underwent 160 lung resection operations at our center. We correlated these results with data from our clinical database, constructed a model predicting the risk of postoperative complications, and assessed its adequacy using receiver operating characteristic curve methodology.
RESULTS: Preexisting cardiovascular disease (p < 0.001), primary lung cancer (p = 0.009), extent of lung resection (p = 0.042), interleukin 6 genotype (p = 0.017), and tumor necrosis factor genotype (p = 0.005) were significantly associated with complications. The odds ratio for complications for rare allele homozygosity was 3.9 (95% confidence interval, 1.4 to 10.4) for interleukin 6 and 15.3 (95% confidence interval, 1.7 to 131.4) for tumor necrosis factor. In multivariate analysis we found that cardiovascular disease (p < 0.001; odds ratio, 4.0 [95% confidence interval, 1.9 to 8.6]), interleukin 6 genotype (p = 0.027; odds ratio, 1.8 [95% confidence interval, 1.1 to 3.1]), and tumor necrosis factor genotype (p = 0.011; odds ratio, 2.5 [95% confidence interval, 1.2 to 5.1]) were independently predictive of complications, with an area under the receiver operating characteristic curve for the entire model of 0.765.
CONCLUSIONS: Carriage of specific alleles, and homozygosity in particular, at loci within the interleukin 6 and tumor necrosis factor genes appears to contribute to the risk of experiencing an adverse event after lung resection.
|
Introduction
|
|---|
Since the first report of pneumonectomy for lung cancer in 1933 [1], there have followed many descriptions of the incidence of mortality and morbidity after major thoracic surgery. The currently reported incidence generally ranges from 3% to 6% for mortality and 15% to 30% for morbidity [2, 3]. Many attempts have been made to predict more accurately those patients who present the greatest risk, with varying degrees of success [2, 4–6]. Previous attempts to characterize those risk factors associated with postoperative complications have focused on environmental and phenotypic characteristics such as cessation of smoking, lung transfer capacity for carbon monoxide [3, 4, 7], and nutritional status [8]. Recent reports suggest that genetic influences may contribute to outcome from coronary artery bypass surgery [9–11], acute respiratory distress syndrome [12], and coronary artery disease [13]. We reviewed the recently published literature on postoperative complications and concluded that there is evidence of a role for inflammation in the cause of cardiopulmonary complications [11, 14–16]. Because genetic variability has been shown to affect the host response to an inflammatory stimulus [17], we formed the hypothesis that variation in loci within genes encoding proinflammatory proteins would, as in other areas of perioperative medicine [13, 18, 19], lead to a higher incidence of postoperative adverse events in patients carrying the proinflammatory alleles. This study was conducted to test that hypothesis.
|
Patients and Methods
|
|---|
Study Population
The population for this study comprised a total of 155 patients with newly diagnosed, histologically confirmed, previously untreated cancer involving the lung who underwent lung resection as part of their treatment. Recruitment of subjects took place between 1995 and 2002 from the new patient pool at The University of Texas M.D. Anderson Cancer Center in Houston, TX. Study staff conducted a computerized review of daily appointment schedules, and new patients were asked to complete a brief eligibility questionnaire. This questionnaire assessed the patient's willingness to participate in the study, smoking status, and prior cancer treatment. All study participants signed an informed consent form approved by the M.D. Anderson Institutional Review Board. After this, patients completed a 1-hour interview with a staff interviewer, after which a 30-mL peripheral blood sample was drawn into coded and heparinized tubes for subsequent analysis.
Surgery and Clinical Data Collection
All patients in this study underwent surgery at our institution between 1995 and 2002: their operative and demographic data are shown in Table 1. Data relating to demographics, clinical risk factors, and postoperative outcome have been collected prospectively for all thoracic surgical patients at our hospital since 1995. These data are entered into the database before, during, and after the operating room visit to ensure timely and accurate recording. Patients undergoing lung resection surgery between the above dates but not recruited into this study did not differ from the study cohort in terms of demographics, outcome, or any other variable that might indicate a selection bias.
Definition of Postoperative Complications
An inflammatory event was defined as any of the following occurring within 30 days of surgery:
- 1 Major pulmonary event as previously described [2] (any or all of the following: acute lung injury, defined as arterial oxygen partial pressure to inspired oxygen fraction ratio less than 300; acute respiratory distress syndrome, defined as arterial oxygen partial pressure to inspired oxygen fraction ratio less than 200; both without evidence of central volume overload; or pneumonia, defined as at least three of the following: leukocytosis greater than 10,000/mm3 or less than 3,000/mm3, temperature greater than 38.5°C or less than 35°C, purulent sputum, persistent infiltrate on chest radiograph, or pathogenic microorganisms from endotracheal aspirate).
- 2 New arrhythmia.
- 3 New thromboembolic disease (deep vein thrombosis confirmed with duplex ultrasound) or pulmonary embolism (high probability ventilation–perfusion scan or visible on computed tomographic angiogram).
- 4 Wound or urinary tract infection (bacteriologically confirmed).
- 5 Acute coronary syndrome (crescendo angina with electrocardiographic changes or troponin-I confirmed myocardial infarction).
- 6 New cerebrovascular event (confirmed by computed tomographic scan).
DNA Extraction and Genotyping
Genomic DNA was isolated from the lymphocytes of whole blood samples using a proteinase-K, sodium dodecyl sulfate, and EDTA-Tris approach. Genotypes for our six candidate loci were determined using the polymerase chain reaction technique, as described in Table 2. All genotypes were determined in triplicate to confirm internal validity, and DD homozygotes at the angiotensin-converting enzyme gene were retyped with different primers to confirm that they were indeed homozygotes and not ID heterozygotes [20].
Statistical Analysis
Univariate comparisons of frequencies were performed using the 
2 test, and comparisons of the means of normally distributed variables were performed using Student's unpaired t test. Comparisons across more than one group (such as in operation type) were performed using analysis of variance with appropriate post hoc testing. In these analyses, a p value of less than 0.05 was used to assess statistical significance, but all variables significant at the 10% level in univariate testing were entered into a logistic regression analysis to build a model for risk prediction. Variables were entered into this model in a stepwise fashion and were designated categorical when appropriate. Variables with a p value less than 0.05 in this model were regarded as being independently associated with complications. The goodness of fit of the entire model was assessed using the Hosmer-Lemeshow statistic, and the ability to discriminate patients who experienced complications from those who did not was measured by calculating the area under the receiver operating characteristic curve of the predicted probabilities. All analyses were carried out using SPSS for Windows version 11.0 (SPSS Inc, Chicago, IL).
|
Results
|
|---|
Clinical Outcome
One hundred fifty-five patients underwent 160 lung resections and had complete data sets and DNA available. Five patients underwent repeat resection during the study period, but the results of the study did not change if these repeat surgeries were excluded; therefore, for completeness data from all 160 operations are reported here. Clinical details, pathologic tumor stage, and operation type are shown in Table 1, along with incidence of cardiovascular comorbidity and preoperative pulmonary function test results. Overall 52 (32.5%) patients experienced the following as a first complication: major pulmonary event, 14 (9%); new arrhythmia, 27 (17%); acute coronary syndrome, 2 (1%); urinary tract infection, 7 (4%); and wound infection, 2 (1%). These were significantly more likely to have occurred in those undergoing anatomic resection (p = 0.042), in those with primary lung cancer (p = 0.009), and in those with preexisting cardiovascular disease (defined as any or all of the following: congestive heart failure, coronary artery disease, cerebrovascular disease, hypertension, peripheral vascular disease, or chronic arrhythmia; p < 0.001). When compared with patients who did not experience a postoperative complication, those who did stayed significantly longer in the hospital (mean ± standard deviation, 6.5 ± 3.4 days versus 10.2 ± 7.5, respectively; p < 0.001; Table 1). They also had a shorter mean duration of survival (31.1 ± 19.1 months versus 25.5 ± 16.7 months, respectively; Table 1), although this did not reach statistical significance (log rank p = 0.063).
Genotype Analysis
Allele frequencies and the results of the 2 test for separation from Hardy-Weinberg equilibrium are shown in Table 3. There was no significant departure from Hardy-Weinberg equilibrium for any of the candidate loci. The distributions of each genotype for patients who did and did not experience a postoperative complication are shown in Table 4. There was a significantly different distribution of genotype for interleukin 6 (p = 0.017) and tumor necrosis factor 
(p = 0.005) between patients who experienced a postoperative complication and those who did not. In both cases the rare allele was associated with a higher incidence of complications, with homozygosity conferring a higher risk than heterozygosity, whereas the more common allele conferred protection against complications (Table 5, Fig 1).
View larger version (11K):
[in this window]
[in a new window]
|
Fig 1. Odds ratios (OR) for complications for individual alleles and homozygous genotypes for interleukin 6 (IL6) and tumor necrosis factor (TNF). (CI = confidence interval.)
|
|
Logistic Regression Analysis
Variables found in univariate analysis to be significantly different between the group with complications and the group without complications at the 10% level were entered into a logistic regression analysis with the purpose of building a model to predict the risk of a complication after lung resection surgery. These variables were age at surgery (p = 0.092), operation type (anatomic versus wedge resection, p = 0.042), preexisting cardiovascular disease (p < 0.001), interleukin 6 genotype (p = 0.017), and tumor necrosis factor genotype (p = 0.005). Details of this model are presented in Table 6. Goodness of fit of the model was tested according to the Hosmer-Lemeshow method [21]. The 2 value for this test was 9.993 for 8 degrees of freedom (p = 0.265), indicating that the model had acceptable fit for the data set. Ability of the model to discriminate those patients likely to experience complications from those unlikely to do so was examined by plotting a receiver operating characteristic curve of the predicted probabilities. This is shown in Figure 2. The area under the curve was 0.77, indicating reasonable discrimination. When the genotype data were removed from the model, the area under the curve reduced to 0.72, which was not statistically different (Fig 2).
View larger version (17K):
[in this window]
[in a new window]
|
Fig 2. Receiver operating characteristic curves of the ability of the multivariate model to predict a postoperative complication with and without inclusion of genotype data. (AUC = area under the curve.)
|
|
|
Comment
|
|---|
We conducted this study to investigate whether variability at loci known to affect inflammatory responses would affect the overall risk of postoperative complications in a high-risk group of patients.
We included only patients undergoing lung resection (as opposed to mediastinal tumor resection or esophagectomy), because we thought this was the best way to measure the magnitude of the surgical insult. Although this can never be exactly quantified, we thought that the following was a reasonable order of increasing severity of insult: single wedge resection, multiple wedge resections, segmentectomy, lobectomy, pneumonectomy. In fact we observed similar rates of events in all these subgroups. Additionally, because of the need for accurate phenotype definition in gene–complex disease association studies, we used strict criteria to determine whether or not a patient experienced a complication.
Our data suggest that variability at inflammatory gene loci may influence the risk of postoperative complications. In both cases (interleukin 6 and tumor necrosis factor ) the rare allele was associated with increased risk. This is in keeping with previous reports of associations of these variants in similar patient populations [13, 18, 19]. However, Gaudino and associates [11, 14, 22] have published several reports of a series of 113 patients in which they demonstrate association of a deleterious outcome with the more common (G) allele of the interleukin 6 locus in question here. This is a feature common to the complex trait–gene association literature and does not necessarily mean either group is wrong. Patient numbers are small in both studies (fewer than 150 patients), and it is possible that the effect is different in different patient populations. This problem will only be resolved with larger studies of well-controlled series of patients and very tight definitions of phenotype.
In our risk prediction model we demonstrated an area under the receiver operating characteristic curve of 0.765, which represents only reasonable discrimination. The area under this curve reduced to 0.72 when the genetic variables were removed, and thus inclusion of the genetic data may appear to make only a small improvement. However, an effect of this size is to be expected if only a very few variants are included. The true effect of genetic variation on postoperative risk likely arises from a combination of multiple variants in multiple genes. Nevertheless, we believe our data do support the concept that genetic variation may be an important, and as yet unrecognized, determinant of postoperative risk. Our results will provide a platform on which to build future iterations of the model that will contain both genetic and environmental variables. There will always be a residual component of risk, comprising random events, human error, and equipment failure among others, and observational models are unlikely to exclude this completely. Rather, a model can identify those patient-related and procedure-related risk factors that may be amenable to intervention. There was no significant difference between survival times by presence or absence of a complication, or by any other of our study variables.
Our choice of candidate loci was governed largely by previous reports of associations with cardiopulmonary disease [12, 18, 23, 24]. There are many cytokines with many variants, and examination of them all is not currently practicable. Genome-wide variant–association studies are still not feasible for humans, and thus the candidate gene approach has become the methodology of choice for this type of research. We selected the variants in this study because of the strength of previously published associations with other similar phenotypes.
Our study has several limitations. First, we cannot offer any functional confirmation of the effects of our genetic variants, inasmuch as we did not measure circulating levels of the protein products of our genes. Second, we did not assess a panel of genomic control variants to rule out the effect of population stratification [25]. However, our data did not suggest any departure from the Hardy-Weinberg equilibrium, which reduces the chance of this problem. Additionally there were no differences in racial background between those who did and did not experience a complication, again making a significant population effect unlikely. Third, we may have made a type one error, finding an association when in fact no such association between the variants reported here and a complications phenotype exists. However, both of our variants were significant in the multivariate model, both were rare allele associations, and both are involved in host innate immunity. We believe that although we cannot exclude the possibility of type one error, this does at least reduce the chance. Last, although clinical events are prospectively entered into our database as they occur, it is possible that we have failed to capture some postoperative complications, and thus have underreported their true incidence, although our results are consistent with those of other centers [26, 27].
In summary, we report an association between inflammatory gene variants and increased risk of postoperative complications. We suggest that these data provide conceptual evidence of a role for host innate immune pathways in the cause of postoperative complications, and that there is a need for larger studies of more candidate genes if the contribution of this risk factor is to be quantified accurately.
|
Acknowledgments
|
|---|
This project was funded through a grant to Drs Shaw and Vaporciyan from the M. D. Anderson Lung Complications Multidisciplinary Research Project.
|
References
|
|---|
- Graham E, Singer J. Successful removal of an entire lung for carcinoma of the bronchus JAMA 1933;101:1371-1374.[Abstract/Free Full Text]
- Vaporciyan AA, Merriman KW, Ece F, et al. Incidence of major pulmonary morbidity after pneumonectomy: association with timing of smoking cessation Ann Thorac Surg 2002;73:420-426.[Abstract/Free Full Text]
- Ferguson MK, Reeder LB, Mick R. Optimizing selection of patients for major lung resection J Thorac Cardiovasc Surg 1995;109:275-283.[Abstract/Free Full Text]
- Walsh GL, Morice RC, Putnam Jr JB, et al. Resection of lung cancer is justified in high-risk patients selected by exercise oxygen consumption Ann Thorac Surg 1994;58:704-711.[Abstract]
- Swisher SG, Deford L, Merriman KW, et al. Effect of operative volume on morbidity, mortality, and hospital use after esophagectomy for cancer J Thorac Cardiovasc Surg 2000;119:1126-1132.[Abstract/Free Full Text]
- Hofstetter W, Swisher SG, Correa AM, et al. Treatment outcomes of resected esophageal cancer Ann Surg 2002;236:376-385.[Medline]
- Ferguson MK, Little L, Rizzo L, et al. Diffusing capacity predicts morbidity and mortality after pulmonary resection J Thorac Cardiovasc Surg 1988;96:894-900.[Abstract]
- Jagoe RT, Goodship TH, Gibson GJ. The influence of nutritional status on complications after operations for lung cancer Ann Thorac Surg 2001;71:936-943.[Abstract/Free Full Text]
- Donahue BS, Byrne DW, Gailani D, George Jr AL. Tissue factor and platelet glycoprotein Ib-alpha alleles are associated with age at first coronary bypass operation Anesthesiology 2003;99:1287-1294.[Medline]
- Donahue BS, Gailani D, Higgins MS, Drinkwater DC, George Jr AL. Factor VLeiden protects against blood loss and transfusion after cardiac surgery. Circulation 2003;107:1003-1008.[Abstract/Free Full Text]
- Gaudino M, Andreotti F, Zamparelli R, et al. The –174G/C interleukin-6 polymorphism influences postoperative interleukin-6 levels and postoperative atrial fibrillationIs atrial fibrillation an inflammatory complication?. Circulation 2003;108(Suppl 1):II-195-II-199.
- Marshall RP, Webb S, Bellingan GJ, et al. Angiotensin converting enzyme insertion/deletion polymorphism is associated with susceptibility and outcome in acute respiratory distress syndrome Am J Respir Crit Care Med 2002;166:646-650.[Abstract/Free Full Text]
- Humphries SE, Luong LA, Ogg MS, Hawe E, Miller GJ. The interleukin-6 –174 G/C promoter polymorphism is associated with risk of coronary heart disease and systolic blood pressure in healthy men Eur Heart J 2001;22:2243-2252.[Abstract/Free Full Text]
- Gaudino M, Di Castelnuovo A, Zamparelli R, et al. Genetic control of postoperative systemic inflammatory reaction and pulmonary and renal complications after coronary artery surgery J Thorac Cardiovasc Surg 2003;126:1107-1112.[Abstract/Free Full Text]
- Chung MK, Martin DO, Sprecher D, et al. C-reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation Circulation 2001;104:2886-2891.[Abstract/Free Full Text]
- Sata N, Hamada N, Horinouchi T, et al. C-reactive protein and atrial fibrillationIs inflammation a consequence or a cause of atrial fibrillation?. Jpn Heart J 2004;45:441-445.[Medline]
- Schwartz DA. The genetics of innate immunity Chest 2002;121(3 Suppl)62–8S.
- Jones KG, Brull DJ, Brown LC, et al. Interleukin-6 (IL-6) and the prognosis of abdominal aortic aneurysms Circulation 2001;103:2260-2265.[Abstract/Free Full Text]
- Schluter B, Raufhake C, Erren M, et al. Effect of the interleukin-6 promoter polymorphism (–174 G/C) on the incidence and outcome of sepsis Crit Care Med 2002;30:32-37.[Medline]
- Fernandez-Llama P, Poch E, Oriola J, et al. Angiotensin converting enzyme gene I/D polymorphism in essential hypertension and nephroangiosclerosis Kidney Int 1998;53:1743-1747.[Medline]
- Hosmer DW, Hosmer T, Le Cessie S, Lemeshow S. A comparison of goodness-of-fit tests for the logistic regression model Stat Med 1997;16:965-980.[Medline]
- Gaudino M, Nasso G, Zamparelli R, et al. Caratterizzazione dell'attivazione infiammatoria e fibrinolitica dopo intervento di bypass aortocoronarico in circolazione extracorporeal (in Italian). [Inflammatory and fibrinolytic activation after coronary artery bypass with extracorporeal circulation] Ital Heart J 2002;3(6 Suppl):646-651.[Medline]
- Lorenz E, Mira JP, Frees KL, Schwartz DA. Relevance of mutations in the TLR4 receptor in patients with gram-negative septic shock Arch Intern Med 2002;162:1028-1032.[Abstract/Free Full Text]
- Hingorani AD, Liang CF, Fatibene J, et al. A common variant of the endothelial nitric oxide synthase (Glu298
Asp) is a major risk factor for coronary artery disease in the UK Circulation 1999;100:1515-1520.[Abstract/Free Full Text]
- Little J, Bradley L, Bray MS, et al. Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations Am J Epidemiol 2002;156:300-310.[Abstract/Free Full Text]
- Wahi R, McMurtrey MJ, DeCaro LF, et al. Determinants of perioperative morbidity and mortality after pneumonectomy Ann Thorac Surg 1989;48:33-37.[Abstract]
- Ferguson MK. Assessment of operative risk for pneumonectomy Chest Surg Clin N Am 1999;9:339-351.[Medline]
- Fernandez-Real JM, Gutierrez C, Ricart W, et al. The TNF-alpha gene Nco I polymorphism influences the relationship among insulin resistance, percent body fat, and increased serum leptin levels Diabetes 1997;46:1468-1472.[Abstract]
- Jahromi MM, Millward BA, Demaine GA. A polymorphism in the promoter region of the gene for interleukin-6 is associated with susceptibility to type 1 diabetes mellitus J Interferon Cytokine Res 2000;20:885-888.[Medline]
- Yoon HJ, Kim H, Kim HL, et al. Interdependent effect of angiotensin-converting enzyme and platelet-activating factor acetylhydrolase gene polymorphisms on the progression of immunoglobulin A nephropathy Clin Genet 2002;62:128-134.[Medline]
This article has been cited by other articles:
|
|
|
|
 
K. J. Hogan, J. K. Burmester, M. D. Caldwell, Q. H. Hogan, D. B. Coursin, D. N. Green, R. M.R. Selzer, T. P. Broderick, D. A. Rusy, M. Poroli, et al.
Perioperative Genomic Profiles Using Structure-Specific Oligonucleotide Probes
Clin. Med. Res.,
September 1, 2009;
7(3):
69 - 84.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. Amar, H. Zhang, B. Park, P. M. Heerdt, M. Fleisher, and H. T. Thaler
Inflammation and outcome after general thoracic surgery
Eur. J. Cardiothorac. Surg.,
September 1, 2007;
32(3):
431 - 434.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. Shaw
Genetics of postoperative complications following thoracic surgery.
Seminars in Cardiothoracic and Vascular Anesthesia,
December 1, 2006;
10(4):
327 - 345.
[Abstract]
[PDF]
|
|
|
|
|
|
|
M. V. Podgoreanu and J. P. Mathew
Prophylaxis Against Postoperative Atrial Fibrillation: Current Progress and Future Directions
JAMA,
December 28, 2005;
294(24):
3140 - 3142.
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. V. Podgoreanu and D. A. Schwinn
New Paradigms in Cardiovascular Medicine: Emerging Technologies and Practices: Perioperative Genomics
J. Am. Coll. Cardiol.,
December 6, 2005;
46(11):
1965 - 1977.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
